|
(11) | EP 3 879 172 A1 |
| (12) | EUROPEAN PATENT APPLICATION |
| published in accordance with Art. 153(4) EPC |
|
|
|
|
|||||||||||||||||||||||||||
| (54) | MACHINE-VISION-BASED INTELLIGENT FOCUSING METHOD FOR MOVING HEAD COMPUTER LAMP |
| (57) An intelligent focusing method of an intelligent moving head light based on machine
vision, which improves an automatic focusing speed and meets the requirements of focusing
sharpness and real-time performance, includes the following steps: S1, prestoring
coordinate positions Z21 and Z22 of the focusing assembly corresponding to stage light projection distances Smax and Smin respectively, and moving the focusing assembly to a position between Z21 and Z22 after the controller receives a command to start an automatic focusing function;
S2, gradually moving the focusing assembly between Z21 and Z22 by a focusing search strategy, calculating a sharpness evaluation value Tn of a current projected image according to an image sharpness evaluation function,
and determining whether Tn is greater than a preset high sharpness threshold TH if yes, completing the focusing, otherwise executing step S3; and S3, determining
whether Tn is less than or equal to Tn-1, if yes, driving the stepper motor by the driver to operate in an opposite direction,
otherwise driving the stepper motor by the driver to operate in an original direction,
and then repeatedly executing step S2 until the sharpness evaluation value Tn is greater than the preset high sharpness threshold TH to complete focusing |
TECHNICAL FIELD
[0001] The invention relates to the field of stage lights, in particularly to an intelligent focusing method of a stage intelligent moving head light based on machine vision.
BACKGROUND
[0002] As an important optical device that provides clear, brilliant and colorful beam effects, an intelligent moving head light is often used in stage scenes, such as concerts, concert halls and studios, where contrasts of lighting effects are strong, backgrounds are complex and changeable, and projection distances change at any time. In a follow spot scene on the stage, it is guaranteed that a beam projected by the light fixture can keep up with a stage object while maintaining imaging sharpness of the light by manual focusing at any time. However, when multiple lights are controlled simultaneously to shine a spotlight on one or more objects, an existing focusing mode has slow focusing speed and is required to improve real-time performance. With developments of performing arts field and digital image processing technology, it would be desirable to realize a fast and accurate intelligent focusing function with clear imaging in intelligent moving head lights.
SUMMARY
[0003] The invention aims to overcome at least one of the above-mentioned deficiencies in the prior art and provides an intelligent focusing method of an intelligent moving head light based on machine vision, which effectively solves problems of poor focusing sharpness and real-time performance when multiple lights are controlled to be follow spotlights at the same time, improves a degree of automation control in stage lighting technologies, and improves quality of performance with lighting.
[0005] An intelligent focusing method of an intelligent moving head light based on machine vision is provided, in which the intelligent moving head light includes a controller, a camera, a driver, a stepper motor and an optical lens assembly that includes a focusing assembly and a zooming assembly. The intelligent focusing method includes the following steps:
S1, prestoring coordinate positions Z21 and Z22 of the focusing assembly corresponding to stage light projection distances Smax and Smin respectively, and moving the zooming assembly to a position between Z21 and Z21 after the controller receives a command to start an automatic focusing function;
S2, gradually moving the focusing assembly between Z21 and Z22 by a focusing search strategy, calculating a sharpness evaluation value Tn of a current projected image according to an image sharpness evaluation function, and determining whether Tn is greater than a preset high sharpness threshold TH, if yes, the focusing step is completed, otherwise executing step S3; and
S3, determining whether Tn is less than or equal to Tn-1, if yes, driving the stepper motor by the driver to operate in an opposite direction, otherwise driving the stepper motor by the driver to operate in an original direction, and then repeatedly executing step S2 until the sharpness evaluation value Tn is greater than the preset high sharpness threshold TH to complete the focusing step.
[0006] For the intelligent moving head lights, after initial conditions of an optical system are determined, the coordinate positions of the zooming assembly and the focusing assembly are fixed. In fact, the reason why a projected image changes from clarity to blurring is actually that a projection distance of the light has changed. This is because a distance between a moving stage object and the light is changing at any time in a follow spot scene on the stage, and the projected image will become blurred when the moving object exceeds a range of depth of field with clear focusing in the last time. In the present invention, the projection distances Smax and Smin are matched to the coordinate positions Z21 and Z22 of the focusing assembly before and after the change thereof in the intelligent moving head light. The focusing assembly coordinate positions are moved between Z21 and Z22 by adopting the focusing search strategy, and whether the focusing step is completed is determined by a comparison result between the sharpness evaluation value Tn and the high sharpness threshold TH of the projected image. When Tn is greater than TH , the focusing step is completed; when Tn is less than or equal to TH, it is determined whether Tn is less than or equal to Tn-1. If Tn is less than or equal to Tn-1. the driver drives the stepper motor to move in a direction opposite to the original movement direction; if Tn is greater than Tn-1, the driver drives the stepper motor to move in the original movement direction. Then, it is determined whether to continue focusing step according to the comparison result between the sharpness evaluation value Tn and the high sharpness threshold TH of the moved projected image until satisfying focusing conditions.
[0007] According to the intelligent focusing method of the present invention, the coordinate positions Z21 and Z22 of the focusing assembly corresponding to the prestored stage light projection distances Smax and Smin are calculated according to the coordinate positions of the zooming assembly and the focusing assembly. In such a way, the focusing assembly is directly positioned within this range, which shortens a focusing route, greatly shortens focusing time, and improves focusing efficiency. An automatic focusing technology and a focusing search strategy of a digital image processing technology are adopted, so that focusing state is determined according to the comparison result between the sharpness evaluation value Tn and the high sharpness threshold TH of the projected image, which can obtain a maximum sharpness value, thereby achieving fine focusing and ensuring focusing accuracy.
[0008] Further, the step S1 also includes:
S11, obtaining a coordinate position Z1 of the zooming assembly in the intelligent moving head light when the controller receives the command to start an automatic focusing function, calculating a coordinate position Z2 of the focusing assembly correspondingly, and driving the stepper motor by the driver to position the focusing assembly to the coordinate position Z2; and
S12, the controller acquiring current projected image data and calculating the sharpness evaluation value Tn of the current projected image according to the image sharpness evaluation function.
[0009] The coordinate position Z1 of the zooming assembly in the intelligent moving head light is obtained when the controller receives the command to start the automatic focusing function, then the coordinate position Z2 of the focusing assembly is calculated correspondingly, and the driver drives the stepper motor to position the focusing assembly to the coordinate position Z2. Since the coordinate position Z2 is a position quite close to an optimal focusing plane, positioning the focusing assembly directly to the Z2 position can achieve rough focusing. Finding an optimal focusing point near the Z2 position can lead to faster focusing. And the controller acquires the current projected image data to calculate and obtain the sharpness evaluation value of the projected image, which provides the foundation for determining the sharpness of the projected image by using the sharpness evaluation value in the following steps.
[0010] Further, the intelligent focusing method also includes:
S4, when the intelligent moving head light works normally, the controller monitoring the coordinate position Z1 of the zooming assembly, and when the position of the zooming assembly changes, repeatedly executing the steps S1 to S3 until the focusing step is completed. For example, the position of the zooming assembly is manually changed, so that a stage light can automatically adjust the position of the focusing assembly to realize automatic focusing and keep the projection clear.
[0011] Further, the intelligent focusing method also includes:
S5, when the intelligent moving head light works normally, the controller monitoring the coordinate position Z1 of the zooming assembly, and determining whether the sharpness evaluation value Tn of the current projected image is greater than or equal to a preset low sharpness threshold TL when the zooming assembly position does not change, if yes, the focusing step is completed, otherwise repeatedly executing step S2 until the focusing step is completed. For example, the rotation of the stage light causes the projection distance to change, the stage light can determine whether to repeatedly execute the step S2 or not according to the comparison result between the image sharpness evaluation function Tn and a low sharpness threshold TL of the current image, that is, whether to continue moving the focusing assembly or not, so that a secondary fine focusing is performed to realize a more accurate intelligent focusing function and guarantee that projected light spots are kept clear.
[0012] Further, the step S11 specifically includes:
acquiring the coordinate position Z1 of the zooming assembly in the intelligent moving head light, and calculating and obtaining the coordinate position Z2 of the zooming assembly through a cam curve of a two-component zoom optical system.
[0013] The present invention actually calculates a movement amount Δ of the focusing assembly by obtaining a movement amount q of the zooming assembly and using the cam curve of the two-component zoom optical system. The optical system of the intelligent moving head light is a mechanical compensation type two-component continuous zoom optical system, which is mainly composed of a front fixed lens, a zooming assembly and a focusing assembly. A specific calculation process of the movement amount Δ of the focusing assembly is that: firstly, setting the coordinate positions of the zooming assembly and the focusing assembly in long focus as an origin, so as to obtain horizontal magnifications m21, m31 and component spacing d231 of the zooming assembly and the focusing assembly. Then according to any movement amount q of the zooming assembly, movement amounts Δ1 and A2 of the focusing assembly can be obtained by using the following calculation formulas.
[0014] It can be seen from the above formulas that, corresponding to any movement amount q of the zooming assembly, two horizontal magnifications m31 and m32 that are reciprocal to each other exist at the same time, and both roots can realize compensation of image plane of system. Therefore, for intelligent moving head lights, when the initial conditions of the optical system are determined, a focus compensation curve of the focusing assembly is determined, that is the cam curve thereof is single. The present invention uses the cam curve of the two-component zoom optical system to calculate the focusing assembly coordinate positions, thereby limiting a search range to allow a starting point of a fine focusing route of the subsequent hill-climbing search method to be closer to the optimal focusing plane, which is beneficial to shorten the focusing time and improve the focusing efficiency.
[0016] The hill-climbing search method is a locally preferred focusing search strategy which starts from the sharpness evaluation value Tn of the current projected image to compare with the sharpness evaluation value Tn-1 of a previous projected image. If Tn is less than or equal to Tn-1. Tn-1 is a maximum value or an optimal focusing position, and the driver drives the stepper motor to operate in the direction opposite to the original movement direction at this time; if Tn is greater than Tn-1, the stepper motor is continuously controlled to continue operating in the original movement direction, so as to climb to the optimal focusing position.
[0017] Further, the image sharpness evaluation function adopts a spectral function, and spectral values of the image adopts a Fourier transform function for calculation, that is,
where n = 0, 1, ..., N-1, m = 0, 1, ..., M-1, N represents rows of a sampled two-dimensional discrete image area N×M, M represents columns of the sampled two-dimensional discrete image area N×M, X(n,m) represents pixel points, and Y(n,m) represents frequency points.
[0018] The present invention adopts the spectral function based on the Fourier transform function, which is currently the most accurate evaluation function for sharpness evaluation. High-frequency components of the Fourier transform correspond to edges of the image, and a focused image always has sharp edges, which means that more high-frequency components are contained. Therefore, the high-frequency contents of the image after the Fourier transform is used as the evaluation function. In an image with clear focusing, a center of a two-dimensional spectrogram is a high-frequency area while the surrounding is a low-frequency area. When there are more high frequency components, spectral data is concentrated in the center, and image focusing is clearer.
[0019] Further, the calculation of the sharpness evaluation value Tm of the current projected image according to the image sharpness evaluation function specifically includes: obtaining a focusing evaluation function according to the spectral values of the Fourier transform spectral evaluation function, that is,
where E{Y(n,m)} represents an average value of the frequency points, σ represents a standard deviation value, namely the sharpness evaluation value Tn of the current projected image.
[0020] The standard deviation value is extracted from a spectral image and is used as the focusing evaluation function to calculate the sharpness evaluation value Tn of the projected image. When the standard deviation value of the image is larger, the high-frequency components of the image occupy a larger proportion, the image is clearer, and the sharpness evaluation value Tn of the projected image is larger. On the contrary, when the standard deviation value of the image is smaller, low-frequency components of the image occupy a larger proportion, the image is blurrier, and the sharpness evaluation value Tn of the projected image is smaller. The present invention uses the focusing evaluation function to calculate the sharpness evaluation value of the image, which avoids a situation that local peaks occur in the Fourier transform spectral evaluation function curve, that is, a calculated optimal sharpness evaluation value is different from an actual optimal sharpness evaluation value, so that sizes of high-frequency components can be compared more accurately to achieve precise focusing.
[0021] Further, a distance between Z21 and Z22 occupies 15% of a maximum route of the focusing assembly.
[0022] After many experiments and tests, a maximum distance value between Z21 and Z22 of the machine vision-based intelligent moving head lights is 0.649 mm, and the maximum route of the focusing assembly is 4.267 mm. It can be seen that in the worst case, a ratio of the distance between Z21 and Z22 to the maximum stroke (0.649/4.267)* 100%=15%, so that the hill-climbing search strategy that is used to limit a search range can avoid searching for invalid position sections, narrow the search range, and improve the search efficiency.
[0023] Further, the maximum value of Smax is less than or equal to 50 m, and the minimum value of Smin is greater than or equal to 3 m.
[0024] Further, a distance in which the focusing assembly moves gradually is from 0.2% to 0.6% of the maximum route of the focusing assembly.
[0025] The distance in which the focusing assembly moves gradually is limited to 0.2% to 0.6% of the maximum route of the focusing assembly, so that the movement distance may not be too small to result in slow focusing, nor too large to miss a sharp focusing point.
[0027] The present invention uses the cam curve of the two-component zoom optical system to calculate the focusing assembly coordinate positions, and then limit the search range and directly locates the focusing assembly to a position close to the optimal focusing plane. In the worst case, the ratio of the distance between Z21 and Z22 to the maximum stroke of the focusing assembly is 15%, which shortens the search range, avoids searching for invalid sections, greatly shortens the focusing time and improves the focusing efficiency.
[0028] The automatic focusing technology of the digital image processing technology is adopted, and the standard deviation value is calculated according to the focusing evaluation function, which avoids a situation that local peaks occur in the Fourier transform spectral evaluation function curve, so that sizes of high-frequency components can be compared more accurately to achieve precise focusing.
BRIEF DESCRIPTION OF THE DRAWINGS
[0029]
FIG. 1 is a flow chart of partial steps according to an embodiment of the present invention.
FIG. 2 is a flowchart of overall steps according to the embodiment of the present invention.
FIG. 3 shows cam curves formed by coordinate positions of a zooming assembly and a focusing assembly with different projection distances.
EMBODIMENTS
[0030] The drawings of the present invention are for illustration purpose only and are not intended to limit the present invention. Some components in the drawings may be omitted, enlarged, or reduced for better illustrating the following embodiments, and sizes of these components do not represent that of actual products. For those skilled in the art, it will be understood that some known structures and descriptions thereof in the drawings may be omitted.
[0031] As shown in FIGS. 1 and 2, the present embodiment provides an intelligent focusing method of an intelligent moving head light based on machine vision, in which the intelligent moving head light includes a controller, a camera, a driver, a stepper motor and an optical lens assembly that includes a focusing assembly and a zooming assembly. The intelligent focusing method includes the following steps:
S1, prestoring coordinate positions Z21 and Z22 of the focusing assembly corresponding to stage light projection distances Smax and Smin respectively, and moving the zooming assembly to a position between Z21 and Z22 after the controller receives a command to start an automatic focusing function;
S11, obtaining a coordinate position Z1 of the zooming assembly in the intelligent moving head light when the controller receives the command to start the automatic focusing function, calculating a coordinate position Z2 of the focusing assembly correspondingly, and driving the stepper motor by the driver to position the focusing assembly to the coordinate position Z2; and
S12, the controller acquiring current projected image data and calculating the sharpness evaluation value Tn of the current projected image according to an image sharpness evaluation function;
S2, gradually moving the focusing assembly between Z21 and Z22 by a focusing search strategy, calculating a sharpness evaluation value Tn of a current projected image according to an image sharpness evaluation function, and determining whether Tn is greater than a preset high sharpness threshold TH, if yes, the focusing step is completed, otherwise executing step S3;
S3, determining whether Tn is less than or equal to Tn-1, if yes, driving the stepper motor by the driver to operate in an opposite direction, otherwise driving the stepper motor by the driver to operate in an original direction, and then repeatedly executing step S2 until the sharpness evaluation value Tn is greater than the preset high sharpness threshold TH to complete the focusing step;
S4, when the intelligent moving head light works normally, the controller monitoring the coordinate position Z1 of the zooming assembly, and when the position of the zooming assembly changes, repeatedly executing the steps S1 to S3 until the focusing step is completed; and
S5, when the intelligent moving head light works normally, the controller monitoring the coordinate position Z1 of the zooming assembly, and determining whether the sharpness evaluation value Tn of the current projected image is greater than or equal to a preset low sharpness threshold TL when the zooming assembly position does not change, if yes, the focusing step if completed, otherwise repeatedly executing step S2 until the focusing step is completed.
[0032] Preferably, the step S11 in the present embodiment specifically includes:
acquiring the coordinate position Z1 of the zooming assembly in the intelligent moving head light, and calculating and obtaining the coordinate position Z2 of the zooming assembly through a cam curve of a two-component zoom optical system.
[0033] Preferably, the focusing search strategy described in the present embodiment is a hill-climbing search method.
[0034] Preferably, the image sharpness evaluation function in the embodiment adopts a spectral function, and spectral values of an image adopts a Fourier transform function for calculation, that is,
where n = 0, 1, ..., N-1, m = 0, 1, ..., M-1, N represents rows of a sampled two-dimensional discrete image area N×M, M represents columns of the sampled two-dimensional discrete image area N×M, X(n,m) represents pixel points, and Y(n,m) represents frequency points.
[0035] Preferably, the calculation of the sharpness evaluation value Tn of the current projected image in the embodiment according to the spectral values of the image specifically includes:
where E{Y(n,m)} represents an average value of the frequency points, σ represents a standard deviation value, namely the sharpness evaluation value Tn of the current projected image.
[0036] Preferably, a distance between Z21 and Z22 in this embodiment occupies 15% of a maximum route of the focusing assembly.
[0037] Preferably, in this embodiment, a maximum value of Smax is less than or equal to 50 m, and a minimum value of Smin is greater than or equal to 3 m.
[0038] Preferably, a distance in which the focusing assembly moves gradually in this embodiment is from 0.2% to 0.6% of the maximum route of the focusing assembly.
[0039] In order to verify the application effect of the intelligent focusing method of the intelligent moving head light based on machine vision, the following three groups of experiments were carried out on the performance thereof.
[0040] Preferably, the experiments used a CMOS camera model MER-500-7UM/UC-L, which has a maximum operating resolution of 2592×1944, about 500,000 pixels, a minimum frame rate of 7 fps, and a pixel size of 2.2 um × 2.2 um. A maximum route of the zooming assembly of the selected intelligent moving head light is about 130 mm, a maximum route of the focusing assembly is about 5 mm, and a resolution of the route is 8 bits, totaling 256 equal parts. An image acquisition resolution of the camera was set to 640*480 to increase an acquisition frame rate to 60 fps; the camera was set to an auto-exposure mode to prevent over-exposure of the acquired image.
[0041] Experiment 1 aimed to verify the image sharpness of automatic focusing. In actual situations, the position of the zooming assembly was fixed for different projection distances, and the same intelligent moving head light was used to carry out automatic focusing test experiments for several times. Experiment 1 selected a projection distance of 4 to 35 meters for testing since such projection range had met most stage application requirements. The results of Experiment 1 are shown in Table 1.
Table 1 Average values of multiple tests at different focusing positions with different
projection distances when fixing zooming assembly position
| Projection distance/m | Focusing evaluation function value | Actual focusing plane position/mm | Auto-focusing focusing plane position/mm | Focusing plane deviation/mm |
| 4.0 | 1.34270 | 4.078 | 4.066 | -0.012 |
| 5.0 | 1.32839 | 3.922 | 3.940 | 0.018 |
| 6.0 | 1.34218 | 3.843 | 3.858 | 0.015 |
| 7.0 | 1.33548 | 3.765 | 3.749 | -0.016 |
| 8.0 | 1.34631 | 3.686 | 3.695 | 0.009 |
| 9.0 | 1.33725 | 3.529 | 3.535 | 0.006 |
| 10.0 | 1.35344 | 3.373 | 3.363 | -0.010 |
| 11.0 | 1.32987 | 3.294 | 3.286 | -0.008 |
| 12.0 | 1.33469 | 3.216 | 3.235 | 0.019 |
| 13.0 | 1.28865 | 3.137 | 3.149 | 0.012 |
| 14.0 | 1.32680 | 3.059 | 3.052 | -0.007 |
| 15.0 | 1.34319 | 2.980 | 2.970 | -0.010 |
| 20.0 | 1.31864 | 2.824 | 2.837 | 0.013 |
| 25.0 | 1.32383 | 2.745 | 2.724 | -0.020 |
| 30.0 | 1.30515 | 2.667 | 2.687 | 0.023 |
| 35.0 | 1.31187 | 2.588 | 2.571 | -0.017 |
[0042] From the experimental data in Table 1, it can be seen that: a maximum deviation between the auto-focusing focusing plane position and the actual focusing plane position was ±0.023 mm, which indicates that the intelligent moving head light based on machine vision in the present invention can meet the requirements of focusing precision and image sharpness.
[0043] Experiment 2 aimed to obtain the cam curve of the two-component zoom optical system by calculating a focusing assembly coordinate position Z2 corresponding to a zooming assembly coordinate position Z1, and limit a search range ΔZ2. From average values of multiple test data, the cam curve as shown in FIG. 3 and the limited search range AZ2 as shown in Table 2 were obtained.
Table 2 Coordinate positions of zooming assembly and focusing assembly and limited
search range
| Zooming assembly coordinate position Z1/mm | 4-meter distance focusing assembly coordinate position Z2_4/mm | 35-meter distance focusing assembly coordinate position Z2_35/mm | Limited search range ΔZ2/mm |
| 0.0 | 0.000 | 0.000 | 0.000 |
| 5.0 | 1.307 | 0.844 | 0.462 |
| 10.0 | 2.347 | 1.724 | 0.622 |
| 15.0 | 3.147 | 2.498 | 0.649 |
| 20.0 | 3.733 | 3.129 | 0.604 |
| 25.0 | 4.062 | 3.556 | 0.507 |
| 30.0 | 4.213 | 3.778 | 0.436 |
| 35.0 | 4.267 | 3.831 | 0.436 |
| 40.0 | 4.222 | 3.813 | 0.409 |
| 45.0 | 4.089 | 3.760 | 0.329 |
| 50.0 | 3.884 | 3.600 | 0.284 |
| 55.0 | 3.636 | 3.360 | 0.276 |
| 60.0 | 3.360 | 3.093 | 0.267 |
| 65.0 | 3.067 | 2.800 | 0.267 |
| 70.0 | 2.756 | 2.507 | 0.249 |
| 75.0 | 2.462 | 2.204 | 0.258 |
| 80.0 | 2.169 | 1.911 | 0.258 |
| 85.0 | 1.867 | 1.627 | 0.240 |
| 90.0 | 1.582 | 1.333 | 0.249 |
| 95.0 | 1.289 | 1.049 | 0.240 |
| 100.0 | 0.987 | 0.756 | 0.231 |
| 105.0 | 0.711 | 0.498 | 0.213 |
| 110.0 | 0.444 | 0.258 | 0.187 |
| 115.0 | 0.222 | 0.062 | 0.160 |
[0044] From the experimental data in FIG. 3 and Table 2, it can be seen that: a maximum value of the limited search range ΔZ2 was 0.649 mm, the maximum route of the focusing assembly was 4.267 mm, so that in the worst case, a ratio of the maximum search range to the maximum stroke of the focusing assembly was: (0.649/4.267)*100%≈15%. Therefore, the ΔZ2 parameter was used as the limited search range of the hill-climbing method, which can avoid searching for invalid position sections, narrow the search range, and improve the search efficiency.
[0045] Experiment 3 aimed to compare the automatic focusing time with and without a cam curve algorithm for several times by using the same intelligent moving head light as to different projection distances in a case where the zooming assembly coordinate position was fixed. The experimental data is shown in Table 3.
Table 3 Average values of multiple tests of automatic focusing time for programs with
and without cam curve algorithm
| Projection distance/m | Automatic focusing time without cam curve/s | Automatic focusing time with cam curve/s | Time-consuming ratio/% |
| 4.0 | 4.1 | 0.7 | 17 |
| 10.0 | 3.9 | 0.5 | 13 |
| 15.0 | 4.0 | 0.6 | 15 |
| 20.0 | 4.2 | 0.9 | 21 |
| 25.0 | 4.5 | 1.0 | 22 |
| 30.0 | 4.3 | 0.9 | 21 |
| 35.0 | 4.4 | 0.8 | 18 |
[0046] From the experimental data in Table 3, it can be seen that: automatic focusing time-consuming ratio with and without cam curve algorithm = (automatic focusing time with cam curve /automatic focusing time without cam curve) × 100%. Therefore, in a projection range of 4 to 35 meters, the time-consuming ratio is about 13 to 22%, that is, the program with the cam curve algorithm takes less time to complete the automatic focusing and has higher focusing efficiency.
[0047] According to the intelligent focusing method of the intelligent moving head light based on machine vision, the cam curve of the two-component zoom optical system is used to calculate a limited search range, and positioning the focusing assembly to a position close to an optimal focusing plane achieves rough focusing; an automatic focusing technology of a digital image processing technology is adopted, and a spectral evaluation function with threshold determination and the hill-climbing search method are selected to achieve fine focusing. Through the above three experiments, it can be seen that: in the projection range of 4 to 35 meters, the maximum deviation between the auto-focusing focusing plane position and the actual focusing plane position was ±0.023 mm, which can meet the requirements of focusing precision and image sharpness; in the worst case, a ratio of the maximum search range to the maximum route of the focusing assembly was about 15%, which shortens the search range and avoids searching for invalid sections; the time-consuming ratio for completing the automatic focusing by the programs with and without cam curve algorithm is about 13 to 22%, that is, the program with the cam curve algorithm takes less time to complete the automatic focusing, and improves the real-time performance of the automatic focusing.
[0048] Obviously, the above embodiments of the present invention are merely examples for clear illustration of the technical solution in the invention, and are not intended to limit the implementations of the present invention. Any modification, equivalent substitution, improvement, or the like within the spirit and principle of the claims of the invention should be included in the scope of the claims of the invention.
1. An intelligent focusing method of an intelligent moving head light based on machine
vision, wherein the intelligent moving head light comprises a controller, a camera,
a driver, a stepper motor and an optical lens assembly that includes a focusing assembly
and a zooming assembly; the intelligent focusing method comprising the following steps:
S1, prestoring coordinate positions Z21 and Z22 of the focusing assembly corresponding to stage light projection distances Smax and Smin respectively, and moving the zooming assembly to a position between Z21 and Z22 after the controller receives a command to start an automatic focusing function;
S2, gradually moving the focusing assembly between Z21 and Z22 by a focusing search strategy, calculating a sharpness evaluation value Tn of a current projected image according to an image sharpness evaluation function, and determining whether Tn is greater than a preset high sharpness threshold TH, if yes, the focusing step is completed, otherwise executing step S3; and
S3, determining whether Tn is less than or equal to Tn-1, if yes, driving the stepper motor by the driver to operate in an opposite direction, otherwise driving the stepper motor by the driver to operate in an original direction, and then repeatedly executing step S2 until the sharpness evaluation value Tn is greater than the preset high sharpness threshold TH to complete the focusing step.
2. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, wherein the step S1 further includes:
S11, obtaining a coordinate position Z1 of the zooming assembly in the intelligent moving head light when the controller receives the command to start the automatic focusing function, calculating a coordinate position Z2 of the focusing assembly correspondingly, and driving the stepper motor by the driver to position the focusing assembly to the coordinate position Z2; and
S12, the controller acquiring current projected image data and calculating the sharpness evaluation value Tn of the current projected image according to the image sharpness evaluation function.
3. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, further comprising:
S4, when the intelligent moving head light works normally, the controller monitoring a coordinate position Z1 of the zooming assembly, and when the position of the zooming assembly changes, repeatedly executing the steps S1 to S3 until the focusing step is completed.
S4, when the intelligent moving head light works normally, the controller monitoring a coordinate position Z1 of the zooming assembly, and when the position of the zooming assembly changes, repeatedly executing the steps S1 to S3 until the focusing step is completed.
4. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, further comprising:
S5, when the intelligent moving head light works normally, the controller monitoring a coordinate position Z1 of the zooming assembly, and when the zooming assembly position does not change, determining whether the sharpness evaluation value Tn of the current projected image is greater than or equal to a preset low sharpness threshold TL, if yes, the focusing step is completed, otherwise repeatedly executing step S2 until the focusing step is completed.
S5, when the intelligent moving head light works normally, the controller monitoring a coordinate position Z1 of the zooming assembly, and when the zooming assembly position does not change, determining whether the sharpness evaluation value Tn of the current projected image is greater than or equal to a preset low sharpness threshold TL, if yes, the focusing step is completed, otherwise repeatedly executing step S2 until the focusing step is completed.
5. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 2, wherein the step S11 specifically comprises:
acquiring the coordinate position Z1 of the zooming assembly in the intelligent moving head light, and calculating the coordinate position Z2 of the zooming assembly through a cam curve of a two-component zoom optical system.
acquiring the coordinate position Z1 of the zooming assembly in the intelligent moving head light, and calculating the coordinate position Z2 of the zooming assembly through a cam curve of a two-component zoom optical system.
6. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, wherein the focusing search strategy is a hill-climbing
search method.
7. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, wherein the image sharpness evaluation function adopts
a spectral function, and spectral values of an image is calculated by a Fourier transform
function, that is,
where n = 0, 1, ..., N-1, m = 0, 1, ..., M-1, N represents rows of a sampled two-dimensional discrete image area N×M, M represents columns of the sampled two-dimensional discrete image area N×M, X(n,m) represents pixel points, and Y(n,m) represents frequency points.
where n = 0, 1, ..., N-1, m = 0, 1, ..., M-1, N represents rows of a sampled two-dimensional discrete image area N×M, M represents columns of the sampled two-dimensional discrete image area N×M, X(n,m) represents pixel points, and Y(n,m) represents frequency points.
8. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 7, wherein the calculation of the sharpness evaluation value
Tn of the current projected image according to the spectral values of the image specifically
includes:
where E{Y(n,m)} represents an average value of the frequency points, σ represents a standard deviation value, namely the sharpness evaluation value Tn of the current projected image.
where E{Y(n,m)} represents an average value of the frequency points, σ represents a standard deviation value, namely the sharpness evaluation value Tn of the current projected image.
9. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, wherein a distance between Z21 and Z22 occupies 15% of a maximum route of the focusing assembly.
10. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, wherein a maximum value of Smax is less than or equal to 50 m, and a minimum value of Smin is greater than or equal to 3 m.
11. The intelligent focusing method of the intelligent moving head light based on machine
vision according to claim 1, wherein a distance in which the focusing assembly moves
gradually is from 0.2% to 0.6% of the maximum route of the focusing assembly.